38 ALASKAN GLACIER STUDIES formed to a sea of crevasses; and not only was it impossible to travel over their surfaces, but it was not even possible to climb up on the glaciers except by the most difficult ice work. Furthermore, the glaciers were even then actively advancing, and the advance extended out even to the fully-stagnant margins, overturning forests of alder and cot-tonwood that were growing on the outer portions of the Malaspina and Atrevida Glaciers. Not only were the ice surfaces broken by a maze of crevasses, and the margins which had hitherto been gently-sloping, moraine-covered icebanks transformed to steep ice cliffs crowned by bristling ice pinnacles, but the margins were pushed forward, and the hither-tof ore stagnant piedmont bulbs were thickened. Haenke Glacier had advanced to tidal condition; the Atrevida and Malaspina Glaciers were pushing into the forest that fringed their margins; and Variegated Glacier had become notably thicker and had crowded out over a rock gorge, destroying the glacial stream that had occupied it in 1905. It was evident from these facts that the glaciers in question had been subjected to some unusual impulse that had caused a sudden forward movement by which they had been pushed forward, thickened, and their surfaces greatly broken. In seeking for an explanation for such a phenomenon, not hitherto recorded, only one cause seemed adequate, namely, the effect of the severe earthquake shocks to which this region was subjected in 1899. The hypothesis put forward by the senior author was that the repeated violent shaking during the earthquakes that occurred between September 3 and 29,1899, threw so much snow and ice into the reservoirs of the glaciers that a wave of motion was started which reached completely down to the terminus of Galiano Glacier before 1905, and which was passing down the four other glaciers during 1906. In testing this hypothesis with the facts available, all were found to be in harmony with it, none were discovered that were opposed to it, and no other hypothesis could be suggested which had no fatal objections to it. While the hypothesis of earthquake cause for this advance seemed, therefore, well supported it \?as desired to subject it to still further test, and it "was one of the main objects of the expedition of 1909 to apply these tests. There were three such tests which we had especially in mind. In the first place, if the advance were due to this cause, it should be confined to the region of violent earthquake shaking. By inquiring regarding the condition of glaciers southeast of Yakutat Bay, as in Glacier Bay, and by study of some of the glaciers of the Prince William Sound region to the northwest we were able to apply this test to some extent, but not so fully as to warrant definite statement of its adequacy in support of the hypothesis. In view of the multitude of glaciers in the region to the southeast and northwest of Yakutat Bay, the fact that among the comparatively few which we were able to study or gain specific information about, we found none that furnished evidence of spasmodic advance, ia only contributory evidence. For final testimony other studies over a wider field are necessary. The second test is that of the behavior of other glaciers in the Yakutat Bay region in the interval between 1906 and 1909. If the hypothesis proposed is correct, probably some of the smaller glaciers of Yakutat Bay had advanced before 1905, and certainly some of the other glaciers of the region ought later to show signs of the appearance of the wave of advance. This was predicted in the report on the expeditions of 1905 and 1906.1 There is reason to believe that there was an advance of some of the smaller glaciers before 1905, though it is now difficult to obtain convincing evidence; but that the i Tarr, R. S., The Yakutat Bay Region, Alaska, Professional Paper 64, U. S. Geol. Survey, 1909, pp. 93-94.